Identification of black zones on national highways: a case study of an Indian city

Abstract

This paper examines the relation between black zone and number of feeder roads joining the four lane, undivided, 8.5 km long National Highway. Primary and secondary data were collected about geometrical features, number of intersection created due to feeder road per kilometer, accident number/frequency and locations. To indentify the black zone on the study stretch, three methods are used such as an accident rate method, accident frequency method and accident severity. After ascertaining a black zone, a technique of negative binomial distribution with gamma distribution and prediction error minimization method was applied for identifying minimum distances in between two feeder road and/or intersection to avoid creation of black zones. The results of this study suggest that feeder network intersection characterized by left turning do in fact play a role in creating black zones other than the traffic flow and geometry. The prediction model suggests that the number of accidents reduces gradually, if feeder road traffic merges with national highway in between minimum spacing should be more than 750 m to 1 km. The performance of the model was successfully tested on the basis of three statistical measures, mean- absolute error (MAE), root mean square error (RMSE) and normalized error (NER). This work suggests that black zone can be prevented through controlling merging of left turning feeder road traffic to national highway traffic by providing minimum in-between safe spacing. Language: en